Dynamic Link Library Loading via Symbol Index Mapping

The patent creates a compact copy of symbol table information by replacing full character string names with integer indices. The global symbol table stores only integer indices instead of complete symbol names, while maintaining the ability to retrieve original symbol names when needed. This copying approach significantly reduces memory consumption while preserving symbol identification functionality.

The patent changes the data type parameter of symbol table entries from character strings to integer indices. This parameter transformation reduces the memory footprint from variable-length strings to fixed-size integers, directly addressing the memory consumption problem while maintaining the indexing and lookup capabilities needed for accurate symbol identification.

The patent performs preliminary actions by pre-computing and storing the mapping relationships between symbol names and integer indices in the global symbol table during the loading phase. This preliminary organization of data allows for efficient O(1) lookup operations later, as the system only needs to retrieve pre-computed index values rather than performing complex string comparisons during execution.

The patent introduces an intermediary mechanism - a hash table or mapping structure - that bridges the relationship between integer indices and symbol names. This intermediary layer allows the system to work with compact integer representations during critical path operations while maintaining the ability to resolve full symbol names when needed, thus reducing complexity in the hot path while preserving completeness.